2005 Special Issue: Learning protein secondary structure from sequential and relational data
Neural Networks - Special issue on neural networks and kernel methods for structured domains
2005 Special issue: Recursive principal components analysis
Neural Networks - Special issue on neural networks and kernel methods for structured domains
Probabilistic based recursive model for adaptive processing of data structures
Expert Systems with Applications: An International Journal
Probabilistic models for melodic prediction
Artificial Intelligence
Classification of graphical data made easy
Neurocomputing
Extracting finite structure from infinite language
Knowledge-Based Systems
A multi-model approach for long-term runoff modeling using rainfall forecasts
Expert Systems with Applications: An International Journal
A Performance evaluation of neural network models in traffic volume forecasting
Mathematical and Computer Modelling: An International Journal
Neural Processing Letters
Computational approaches to sentence completion
ACL '12 Proceedings of the 50th Annual Meeting of the Association for Computational Linguistics: Long Papers - Volume 1
A challenge set for advancing language modeling
WLM '12 Proceedings of the NAACL-HLT 2012 Workshop: Will We Ever Really Replace the N-gram Model? On the Future of Language Modeling for HLT
Learning sequence neighbourhood metrics
ICANN'12 Proceedings of the 22nd international conference on Artificial Neural Networks and Machine Learning - Volume Part I
Computer Speech and Language
Can we build language-independent OCR using LSTM networks?
Proceedings of the 4th International Workshop on Multilingual OCR
Selective Recurrent Neural Network
Neural Processing Letters
Deep learning of representations: looking forward
SLSP'13 Proceedings of the First international conference on Statistical Language and Speech Processing
Proceedings of the Fourth Symposium on Information and Communication Technology
Pattern Recognition Letters
Hi-index | 0.01 |
Recurrent neural networks can be used to map input sequences to output sequences, such as for recognition, production or prediction problems. However, practical difficulties have been reported in training recurrent neural networks to perform tasks in which the temporal contingencies present in the input/output sequences span long intervals. We show why gradient based learning algorithms face an increasingly difficult problem as the duration of the dependencies to be captured increases. These results expose a trade-off between efficient learning by gradient descent and latching on information for long periods. Based on an understanding of this problem, alternatives to standard gradient descent are considered